Return to BSD News archive
Path: sserve!manuel!munnari.oz.au!uunet!cis.ohio-state.edu!pacific.mps.ohio-state.edu!zaphod.mps.ohio-state.edu!sdd.hp.com!spool.mu.edu!agate!soda.berkeley.edu!wjolitz From: wjolitz@soda.berkeley.edu (William F. Jolitz) Newsgroups: comp.unix.bsd Subject: Re: [386bsd] Strange solution to booting problem Date: 17 Sep 1992 19:48:26 GMT Organization: U.C. Berkeley, CS Undergraduate Association Lines: 48 Message-ID: <19aneaINNkp@agate.berkeley.edu> References: <1992Sep17.004545.2065@drycas.club.cc.cmu.edu> NNTP-Posting-Host: soda.berkeley.edu In article <1992Sep17.004545.2065@drycas.club.cc.cmu.edu> ghod@drycas.club.cc.cmu.edu writes: >Greetings.... > > I'd like to share with you a strange experience I just had concerning >the dist.fs bootable image. I've been trying to get a 3.5" copy of it to boot >on my 386 SX for the past couple of days with no success. Each time I try to >boot, the system hangs almost immediately. At best, it only reads the boot >sector, then it locks up tight. The only observable activity prior to the >system freezing is the momentary blinking of the keyboard LEDs. After that, >nothing. Zip. Nada. Even the RESET button won't work. > > ...[tried everything else to no avail]... > > Then tonight, a few minutes ago in fact, I got the standard dist.fs >to boot. Here's what happened: I have a pile of 3.5" 1.44meg floppies lying >around (none of which are labeled, of course) which will soon hold the 386bsd >binary distribution set. They're all MS-DOS formatted, but not bootable. >Somehow I got confused and tried to boot off one of them by mistake, resulting >in the familiar "Non-system disk" error message being displayed. So I replaced >it with the dist.fs disk and hit RETURN, fully expecting the machine to lock up >yet again. Only this time, it worked like a charm. I tried this trick, which I >now call 'the bait and switch', repeatedly and it worked each time. I'll be >darned if I can figure out what's going on, but I'm hoping someone in this >discussion can shed some light on the subject. ... Ah ha! This is actually very significant! Good for you (and lady luck). Some BIOS use slightly different parameters for the floppy disk controller to make it faster -- if this fails, it should fall back on the "less efficient" values. What I believe happened is this: By inserting the wrong non-bootable floppy, a program left by DOS was run (hence, your error message). which adjusted the values to "slow" When you immediately followed up with the Tiny 386BSD floppy, it worked. In other words, you imposed a timing delay in the bootstrap (it must be loading too quickly). Simply trying to boot the floppy over and over could not "readjust" the values. Here's a simple test to validate this -- measure the time between the blink of the floppy disk light and the keyboard blinks for the failed case and the successful case. It should lengthen. This is an important find. The bootstrap code will have to be carefully reexamined (as usual). Lynne Jolitz.